Re: Linux AX25 KISS driver for Kenwood D700

[Ralf Baechle DL5RB <ralf@linux-mips.org>]

On Thu, Dec 17, 2009 at 02:34:47PM -0800, Bill V WA7NWP wrote:

> Is this where the maintainers of the Linux KISS code would be found?

Generic way of finding a maintainer for a piece of kernel code:

 o look in the MAINTAINERS file in the root directory of the kernel source
   tree
 o Look for email addresses in the source code in question itself.  That
   would have yielded my name and address.
 o If still unsuccessful, look through the git source tree.
 o If you don't have git, dig it up on git web at http://git.kernel.org;
   in this particular case
   http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=history;f=drivers/net/hamradio/mkiss.c;h=7db0a1c3216c7b8dc86703ac23074b9afbe82867;hb=6b7b284958d47b77d06745b36bc7f36dab769d9b
   which will yield a list of people who last modified the the code or
   approved changes to it - that are the Signed-off-by:, Reviewed-by:,
   Tested-by: etc. headers at the bottom of commit messages.

>  A simple tweak would eliminate the possibility of some ASCI sequences
> kicking the Kenwood D7 or D700 radios out of KISS mode...
> 
> We really need to feed this info back to Kenwood so they correctly fix
> the D710.   I don't hold any hope in firmware updates for the D7/D700
> so it seems tweaking the KISS generator is the best answer.

I agree.

Kenwood has a tradition of shipping substandard TNC implementations, see
for example the TNC in the otherwise nice TH-D7 handheld.

> The 710 “bug” is not a violation of the KISS protocol…it is the result
> of trying to use the same serial port for both a TNC and a Front panel
> control. The problem is when in KISS mode (which sends binary packets
> “encapsulated in the KISS protocol”  it is possible to have the
> character sequence “TC 0<Cr>”  as binary or  character data within the
> KISS encapsulation. This sequence will be interpreted by the serial
> port controller in the Control head of the D710 as “escape to control
> mode” and which will switch the serial interface from TNC mode to
> front panel control mode.  This can have disastrous affects when
> trying to transfer binary files using KISS.

If you can't speak normal KISS to the interface I very well call it a 
violation of the KISS protocol.  And it's a terribly, terribly lame one.
Reminds me of the modems taking "+++" in the data stream for a command 
but those vendors at least had the excuse of Hayes' patents.

> One solution would be to entice Kenwood to modify the firmware in the
> control head to having entered the KISS protocol to ignore all mode
> control commands including “TC 0<Cr>”. This would require using the
> standardized escape KISS sequence {0xC0, 0xFF, 0xC0} prior to any
> control panel commands and then re enabling KISS to continue TNC
> operation…This gets tricky and is very likely to impact existing
> programs that now work.

The solution which Hayes has a patent on was to only accept the +++ escape
to command mode sequence if it is framed by a short delay.  Thankfully
Wikipedia has an article coverting this in details at
http://en.wikipedia.org/wiki/Time_Independent_Escape_Sequence

Note the expiry date of the patent so the approach should be possible for
a modem even in the united patent hell states of america.  Still a workaround
for existing firmware is needed.

> The solution we used and which appears to work well and is solid is to
> escape the “C” character.  So when any data is sent to the TNC in KISS
> mode if there is an ASCII “C” it is replaced by the sequence <0xDB>”C”
> which essentially escapes the “TC 0<Cr>” sequence. The KISS protocol
> (in the D710’s KISS TNC implementation) removes the <0xDB> which is
> the frame escape so no modification of the transmitted data stream is
> actually made. Thanks to Peter Woods for figuring this out.

It would be easy enough to add this to the escaping code in the mkiss.c
driver in the kernel - but I can't test this.

  Ralf
--
To unsubscribe from this list: send the line "unsubscribe linux-hams" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html